Re: We need an architecture, not finger pointing.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



> On Oct 28, 2015, at 4:28 PM, Ned Freed <ned.freed@xxxxxxxxxxx> wrote:
> > In any case, the bottom line is you're jumping to conclusions about a vast
> > infrastructure that's currently handling in excess of a trillion messages a day
> > based on what appears to be little more than your personal experience. And I
> > must say if your experience is what you say it is, it's more than a little
> > atypical.

> Er, the big mail services mostly just silently drop mail they don’t like,
> as far as I can tell,

Actually, most of the big providers return a 5yz error, and they do it
as early in the SMTP dialogue as possible. If the message is accepted it
generally gets delivered, although it may be marked as spam.

A few providers seem to think that its a bad idea to say that a message
has been dropped as spam, but this tends to be more of a low end thing.

Things like size limit or bad content are almost always handled with an
outright rejection.

> and they don’t tend to do policies like we’re
> discussing here:

That depends on the provider. Some offer essentially no filtering capability;
others offer fairly extensive configurable server side filtering.

> they just evaluate how likely the mail is to be something the
> reader wants to see, and deliver it or junk it on that basis.   You are right
> that they do a very good job of it, but it’s quite an expensive job, and
> that’s part of my criticism: it need not be nearly so expensive, and it need
> not be something only big companies can do.

Actually, there are any number of reasonably priced AS/AV solutions for MS
Exchange class mail systems. (Admittedly this may change as more and more stuff
shifts to the cloud.)

The selection for non-Exchange systems is smaller, but there are still options
available. For my small SOHO-class mail system, I pay around $100 a year. This
gets me service for a handful of accounts that's roughly comparable in quality
to what the big players offer. (It took a couple of hours to get it set up and
configured, but most of that had to do with getting things running on a Mac
rather than a PC or Linux machine than it did with any mail system issues.)

I don't think that's terribly expensive, but YMMV.

> This conversation is specifically about a problem with IETF mailing lists, so
> the points I was making do apply here.   I am in fact curious to hear if the
> big mail providers are doing things like what I described in the previous
> message—if so, that’s cool in principle.   But if it’s not described in
> the RFCs, you haven’t actually contradicted the point I was making, which is
> that it ought to be reflected in the standards.

AFAIK everything I've talked about is described in an RFC somewhere. Indeed,
the problem isn't that it's not there, it's that there's too much and it's
difficult for inexerienced people to match up the right capabilities and
architecture to solve a given problem.

				Ned




[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]